Skip to main content
ARS Home » Research » Publications at this Location » Publication #188855

Title: DEPENDABILITY AND EFFECTIVENESS OF SEASONAL FORECASTS FOR AGRICULTURAL APPLICATIONS

Author
item Schneider, Jeanne
item Garbrecht, Jurgen

Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/1/2006
Publication Date: 11/1/2006
Citation: Schneider, J.M., Garbrecht, J.D. 2006. Dependability and effectiveness of seasonal forecasts for agricultural applications. Transactions of the ASABE. 49(6):1737-1753.

Interpretive Summary: NOAA's Climate Prediction Center issues seasonal climate forecasts predicting total precipitation and average air temperature for three-month periods out to a year in advance. The utility of these seasonal forecasts for agricultural applications will depend on forecast characterisitcs, including dependability and effectiveness. Dependability is defined as the ability of the forecasts to correctly predict future climate probabilities. Effectiveness refers to the ability of the forecasts to provide information beyond that achieved by standard climatological considerations. Measures of dependability and effectiveness are developed for the NOAA probability of exceedance forecasts, and evaluated for the forecasts issued for 1997 through early 2005 for the contiguous United States. Results depend on forecast variable, direction of forecast (wetter/drier, warmer/cooler), season, and forecast lead time. Forecasts for warmer than average conditions are frequent and dependable for most of the U.S.; forecasts for cooler than average conditions are infrequent and rarely dependable. Forecasts for wetter than average conditions are less frequent and dependable than temperature forecasts; with the exception of Arizona and Florida, forecasts for drier than average conditions are not as frequent or dependable as the wetter forecasts. Overall, in the Desert Southwest, southern and eastern Texas, the Gulf Coast, Florida, and parts of the Pacific Northwest, temperature forecasts have relatively high effectiveness, primarily in November through July, at all lead times. In the same regions, precipitation forecasts have moderate effectiveness in October through February, for lead times up to 6.5 months. For the rest of the U.S., only temperature forecasts show any effectiveness.

Technical Abstract: NOAA's Climate Prediction Center issues seasonal climate forecasts predicting total precipitation and average air temperature for three-month periods out to a year in advance. The utility of these seasonal forecasts for agricultural applications will depend on forecast characterisitcs, including dependability and effectiveness. Dependability is defined as the ability of the forecasts to correctly predict future climate probabilities. Effectiveness refers to the ability of the forecasts to provide information beyond that achieved by standard climatological considerations. Measures of dependability and effectiveness are developed for the NOAA probability of exceedance forecasts, and evaluated for the forecasts issued for 1997 through early 2005 for the contiguous United States. Results depend on forecast variable, direction of forecast (wetter/drier, warmer/cooler), season, and forecast lead time. Forecasts for warmer than average conditions are frequent and dependable for most of the U.S.; forecasts for cooler than average conditions are infrequent and rarely dependable. Forecasts for wetter than average conditions are less frequent and dependable than temperature forecasts; with the exception of Arizona and Florida, forecasts for drier than average conditions are not as frequent or dependable as the wetter forecasts. Overall, in the Desert Southwest, southern and eastern Texas, the Gulf Coast, Florida, and parts of the Pacific Northwest, temperature forecasts have relatively high effectiveness, primarily in November through July, at all lead times. In the same regions, precipitation forecasts have moderate effectiveness in October through February, for lead times up to 6.5 months. For the rest of the U.S., only temperature forecasts show any effectiveness.